Global professional technology services company, Cognizant has launched its new Neuro AI Multi-Agent Accelerator and Multi-Agent Service Suite.
In contrast to its long product name, this new software tool is essentially a no-code development framework offering a selection of pre-built reference network topologies, templates and architectures to help organizations build agentic AI functions.
Why Agentic AI?
Rapidly rising out of the initial fires that have fueled the development of generative AI and the now-popularized extended use of retrieval augmented generation (RAG) custom-aligned domain-specific extensions, interest in agentic AI has (arguably) carried the torch forward as the vanguard movement in the automation intelligence market.
Agentic AI has garnered special interest for its ability to deliver AI agents, which are composite deployments of AI code designed to exhibit human-like intelligence, that work towards specified goals with little to no human intervention. Agentic AI learns as it works and therefore, when all AI model factors are in balance, becomes more adept at its prescribed tasks over time.
But agentic AI is complex, as a result tough to build and even tougher to codify and standardize. With some 350,000 technical practitioners in its ranks, Cognizant thinks it can draw upon its resources and experience to help provide a means to rapidly prototype, customize and scale multi-agent systems.
Vishal Gupta, partner in the data and AI group at the technology analyst house, Everest Group, suggests Cognizant has answered that clarion call to enable organizations to integrate agents into their IT stack infrastructure effectively.
“The rise of autonomous agent networks in enterprise workflows underscores the urgent need for a structured framework enabling seamless interaction and coordination among agents,” said Gupta. “Cognizant tackles this challenge head-on, with a multi-agent development framework that delivers a solution laser-focused on scalability and interoperability.”
Beyond Plain Old Fixed Automation
Cognizant says its Neuro AI service will help change business processes. Using AI agents for adaptive operations, real-time decision-making and personalized customer experiences across business functions including IT and finance, and onward to sales and marketing, it hopes to normalize and standardize the implementation of agents.
The company feels that traditional workflows and “fixed” automation no longer meet customer expectations because of escalating costs and demand for real-time adaptability. Agentic AI intelligence is the more “adaptive” form of automation.
“AI agents are transforming enterprise operations through task automation and reducing manual effort, enabling employees to focus on strategic activities,” said Babak Hodjat, CTO of AI at Cognizant. “However, without collaboration among specialized agents, software systems will remain disconnected from larger business goals. Neuro AI Multi-Agent Accelerator and the Multi-Agent Service Suite allow clients to quickly build and deploy agents into the fabric of their organization, so that they work together across entire businesses to assist humans in lots of roles, from finance and IT to marketing and sales.”
Hodjat further explains that pre-built multi-agent network templates provide a starting point for organizations to address enterprise functions like sales and marketing, finance and investor relations. Further still, agentic AI in this form, can be put to task on workflows related to industry-specific processes including supply chain management, customer service and, as an additional example, insurance underwriting.
“Additional agent networks can be rapidly created using natural language descriptions to fit different scenarios and client use cases and expanded to include third-party agents,” said CTO Hodjat. “While single agents are good at very focused tasks, they can be susceptible to hallucinations or falling into tailspins when faced with more complex problems. With Neuro AI Multi-agent Accelerator and Services Suite, companies can now break up complex tasks between different, specialized agents and more easily connect to third-party agent networks for greater productivity and process automation.”
Neuro AI Toolset Functions
In terms of the actual functions on offer here, Cognizant details an accelerated agentification process through the provision of pre-built multi-agent networks that offer customizable building blocks incorporating industry and functional best practices. A business can start with proven AI agent networks to reduce implementation time and technical risks. Customizing these networks allows for differentiation and adaptation to unique client needs while maintaining the speed and reliability of the pre-built network.
Also included is the ability to integrate newly developed multi-agent networks with pre-existing and third-party agentic systems using simple APIs. Users can perform encapsulation of agent responsibilities to allow for extensibility and automatic routing of tasks to the right AI agents. Ambiguity resolution over the network of agents allows new agents to be added more easily while minimizing errors and improving response times. Cognizant has also enabled data engineering teams, deploying its service to manage large workloads by distributing tasks across multiple servers.
Neuro AI Multi-Agent Accelerator is part of the Multi-Agent Services Suite, a broader suite of services designed to offer a standardized approach to agentic AI functions and manage them in production with enhanced security and compliance.
“The Cognizant Neuro AI Platform is a multi-agent orchestration system designed to streamline creating, deploying and managing AI solutions across diverse industries. The platform’s GUI-first approach and focus on real-world utility empower non-technical users to lead AI initiatives, making it a strong contender in AI orchestration,” explained Kiumarse Zamanian, founder & principal, Accel Experts. “It is differentiated by its business-centric design, i.e. unlike developer-focused platforms. This platform caters to business leaders with an intuitive GUI and simplified workflows. It addresses the entire AI lifecycle with industry-specific templates for streamlined deployment; and it’s [ability to offer] synthetic data for testing as a standout feature for rigorous testing.”
Decentralized Decision-Making
Multi-agent solutions go beyond single agents by enabling decentralized decision-making, where agents act independently, yet, collaboratively to solve complex, interdependent problems. They are designed to provide scalability across functions and geographies, and allow expansion without overhauling systems. Additionally through redundancy, they offer resilience ensuring continuity even if individual agents fail.
